Saylor has been around long enough that a lot of people do not realise it still exists. Michael Saylor, yes, the bitcoin and MicroStrategy one, set up the Saylor Foundation in 1999 and started publishing free college-level courses in 2008. It became Saylor Academy, and in March 2026 Florida's Commission for Independent Education approved it to operate as Saylor University. It now runs a small set of business master's programmes alongside the free courses, holds international accreditation from ASIC and says it is pursuing US accreditation.
None of that changes the free course library, which is still there, still free, and still around 300 courses wide. CS205 Building with Artificial Intelligence is the current AI course. It runs across ten units: what AI is, the agent-based approach, why machine learning matters, machine learning algorithms, problem-solving methods, search algorithms, iterative improvement such as hill climbing and simulated annealing, game-playing models, natural language processing including some generative AI, and reasoning agents. Saylor estimates 48 hours. Only the final exam counts.
You need 70 percent, you get three attempts with a 14-day wait between each, and there is a practice exam you can take as many times as you like. Passing gets you a certificate with 4.8 CEUs on it. If that syllabus looks familiar, it is because it follows the shape of the standard undergraduate AI course that has been taught from Russell and Norvig for decades. I actually like that. Most of the AI courses I review now start at ChatGPT and stay there, and people come out of them unable to explain what a search problem or an agent is.
CS205 gives you that vocabulary, which helps a lot when you later read about planning, agents and tool-using LLMs. The weakness is the Saylor model itself. Saylor does not film lectures or write much original material. It curates open textbooks, university lecture videos and articles, adds a short framing paragraph, and sets quizzes. That keeps it free, but it also means you are reading a different author every few pages, and the overall course has no teacher personality at all.
The course asks you to program in a modern language with readily available libraries but does not really make you do it, so you can pass the exam without writing code. I would treat the programming as mandatory anyway and implement the search algorithms yourself. On credentials, be realistic. The certificate is free and unproctored. It is a fine line on a LinkedIn profile and a useful record for yourself, but no employer is going to treat it like an exam they have heard of.
Saylor has historically offered proctored exams and ACE or NCCRS credit recommendations on some courses, so if you need credit, check the course page for a current credit option before you assume one exists. My recommendation is to use CS205 as the theory half of a two-part self-study plan. Do it alongside something hands-on, like Kaggle Learn or fast.ai, and you get both the concepts and the practice for free.